Question

I have data in this format (tab separated columns)…

Name_00001  Annotation_1    gene    329412  330509  .   -   .   ID=TAXON_00029;Name=TAXON_00029
Name_00001  Annotation_1    mRNA    329412  330509  .   -   .   ID=TAXON_00029T0;Parent=TAXON_00029
Name_00001  Annotation_1    exon    193957  194252  .   +   .   ID=exon:TAXON_00022T0:1;Parent=TAXON_00022T0
Name_00001  Annotation_1    exon    195500  197118  .   +   .   ID=exon:TAXON_00022T0:2;Parent=TAXON_00022T0

…and I would like to sort by the first 5 numbers in column 9 in ascending order. I.e. so the file looks like this:

Name_00001  Annotation_1    exon    193957  194252  .   +   .   ID=exon:TAXON_00022T0:1;Parent=TAXON_00022T0
Name_00001  Annotation_1    exon    195500  197118  .   +   .   ID=exon:TAXON_00022T0:2;Parent=TAXON_00022T0    
Name_00001  Annotation_1    gene    329412  330509  .   -   .   ID=TAXON_00029;Name=TAXON_00029
Name_00001  Annotation_1    mRNA    329412  330509  .   -   .   ID=TAXON_00029T0;Parent=TAXON_00029

I have tried to use Unix or Perl, but my skills are very limited…

Answer 1

If you really want to sort by the first 5 digits in column 9,
that is something you cannot literally explain to sort, but as you suggested, perl is more flexible:

perl -e 'sub key { 0+((split(/\s+/,$_[0]))[8] =~ /(\d{5})/)[0] }; print sort { key($a) <=> key($b) } <>' < unsorted > sorted

This assumes the five digits always exist and are always consecutive but you can take it from there.
It’s better to put the code into a file, add use warnings; use strict; and increase legibility:

#!/usr/bin/env perl

use strict;
use warnings;

sub key {
    my $line = $_[0];
    my @fields = split(/\s+/,$line);
    my ($keystring) = $fields[8] =~ /(\d{5})/;
    my $key = 0 + $keystring;
    $key
}

print sort { key($a) <=> key($b) } <>